Bug 100829 – [EFL] CMake shows ENABLE_3D_RENDERING and ENABLE_WEBGL is still OFF when AC is enabled

• Status: RESOLVED
• Resolution: FIXED
• Component: WebKit EFL

[Joone Hur]

When tiled backing store is turned on, graphics acceleration stuff is enabled accordingly as follows:

http://trac.webkit.org/changeset/132571
http://trac.webkit.org/changeset/132627

However, CMake still shows that ENABLE_3D_RENDERING and ENABLE_WEBGL is still OFF when running build-webkit with --tiled-backing-store.

[Joone Hur]

Created attachment 171598 [details]
patch

[Raphael Kubo da Costa (:rakuco)]

Comment on attachment 171598 [details]
patch

Simply moving the call to WEBKIT_OPTION_END() to the end of the file in OptionsEfl.cmake should solve the problem in a cleaner way.

[Joone Hur]

(In reply to comment #2)
> (From update of attachment 171598 [details])
> Simply moving the call to WEBKIT_OPTION_END() to the end of the file in OptionsEfl.cmake should solve the problem in a cleaner way.

I tried, but it doesn't work.

[Raphael Kubo da Costa (:rakuco)]

(In reply to comment #3)
> (In reply to comment #2)
> > (From update of attachment 171598 [details] [details])
> > Simply moving the call to WEBKIT_OPTION_END() to the end of the file in OptionsEfl.cmake should solve the problem in a cleaner way.
> 
> I tried, but it doesn't work.

That's probably because of the interaction between normal variables and cache variables in CMake.

SET(ENABLE_WEBGL 1) in OptionsEfl.cmake sets a normal variable, and later the call to OPTION(ENABLE_WEBGL ...) will set a cache variable with the same name and remove the original normal variable.

I think the cleanest solution here is to move the call to WEBKIT_OPTION_END() to the end of OptionEfl.cmake and modify the macro so that the initial value it uses in the OPTION() call is the value of the variable itself if it is defined, and the initial value otherwise. This would still require you to clean up your cache nonetheless, as CMake by default won't change the options you already have in your cache when you call a command such as OPTION().

Patrick, what do you think?

[Raphael Kubo da Costa (:rakuco)]

*** Bug 102903 has been marked as a duplicate of this bug. ***

[Halton Huo]

How about my patch for the dup bug 102903? it works as wish. could somebody assign this bug to me so that i can add upload patch.

[Halton Huo]

Created attachment 175555 [details]
Patch

[Halton Huo]

Anyone could review? Joone, Gyuyoung?

[Gyuyoung Kim]

Comment on attachment 175555 [details]
Patch

Hmm, WebKit doesn't like that a bug lands multiple patches. I think you guys need to merge these patches into a patch.

[Halton Huo]

Kim, as joone said on comment #3, first patch does not work. The second one I uploaded works well, not need with first patch.

[Gyuyoung Kim]

Comment on attachment 175555 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=175555&action=review

I think rakuco might to take a final look.

> ChangeLog:4
> +        AC is enabled

It looks this is unneeded line change.

[Halton Huo]

Created attachment 175634 [details]
Patch

[Halton Huo]

(In reply to comment #12)
> Created an attachment (id=175634) [details]
> Patch
Update patch addressing Gyuyoung's comment at #11

[WebKit Review Bot]

Comment on attachment 175634 [details]
Patch

Clearing flags on attachment: 175634

Committed r135813: <http://trac.webkit.org/changeset/135813>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.